If anyone is still doing topic modelling, I'd love some user feedback (thanks to @alanyliu.bsky.social for some initial suggestions). This was a fascinating example of how we can quickly adapt and update existing open-source tools. I hope to develop it further.

Are you going to continue to develop the app, or is it functionally complete? I'm curious as a user, but I think this question is important for these types of quickly spun up tools, as you hinted in your earlier post.

Looking forward to seeing where this goes!

Interesting. I'm currently working on the MALLET interface for Lexos and have been using Claude 3.5 to help with test functions (with limited success). But the lesson here is that we need to start thinking about AI as a (the?) potential user when we design our software...

It's sounded like armageddon here since late last night.

Also, the types of projects we can envision are different when a single course release (reducing us to three courses per semester) can blow nearly the entire grant budget. Even if we are guaranteed to get the grant, at some point it becomes not worth it.

Not sure if it's quotable for your purposes, but Gwyneth Glyn's "Cofia fi at" is lovely.

We might need some brainstorming about success strategies for the CSU. Let's put it on the agenda for @dhatcsu.bsky.social.

Hope to catch you there!